As part of a 38-laboratory international study individual laboratory stocks of five Salmonella typhimurium tester strains were compared with reference strains in their response to a mutagen, 4-nitroquinoline-N-oxide (4NQO). The results from all laboratories were analyzed in order to determine the levels of agreement within and between laboratories for each Salmonella strain. It was concluded that "genetic drift" of the tester strains was not a significant factor in interlaboratory variability. Individual laboratory techniques and test conditions appear to be the major sources of laboratory-to-laboratory variability.